The interplay between chemistry and transport phenomena during the fast pyrolysis of cellulose, lignin and biomass

Hele tekst

(1)The Interplay Between Chemistry and Transport Phenomena During the Fast Pyrolysis of Cellulose, Lignin and Biomass. Pushkar Satish Marathe.

(2)

(3) THE INTERPLAY BETWEEN CHEMISTRY AND TRANSPORT PHENOMENA DURING THE FAST PYROLYSIS OF CELLULOSE, LIGNIN AND BIOMASS. Pushkar Satish Marathe.

(4)

(5) THE INTERPLAY BETWEEN CHEMISTRY AND TRANSPORT PHENOMENA DURING THE FAST PYROLYSIS OF CELLULOSE, LIGNIN AND BIOMASS. DISSERTATION. to obtain the degree of doctor at the University of Twente, on the authority of the rector magnificus, prof.dr. T.T.M. Palstra, on account of the decision of the Doctorate Board, to be publicly defended on Friday the 13th of September 2019 at 16:45 hours. by. Pushkar Satish Marathe. born on the 16th August 1988 in Mumbai, India. iii.

(6) This dissertation has been approved by: Supervisor: prof.dr. S.R.A. Kersten. The research described in this thesis was conducted in the Sustainable Process Technology (SPT) group at the University of Twente, The Netherlands. This work is financially supported by the Netherlands Organisation for Scientific Research (NWO); Project number - 717-014-006.. Cover design: Prachi P. Buche Marathe and Pushkar S. Marathe Printed by: Gildeprint Lay-out: in LATEX by Pushkar S. Marathe ISBN: 978-90-365-4845-8 DOI: 10.3990/1.9789036548458. © 2019 Pushkar S. Marathe, The Netherlands. All rights reserved. No parts of this thesis may be reproduced, stored in a retrieval system or transmitted in any form or by any means without permission of the author.. iv.

(7) GRADUATION COMMITTEE: Chairman: Secretary:. prof.dr J.L. Herek prof.dr J.L. Herek. University of Twente University of Twente. Supervisor:. prof.dr. S.R.A. Kersten. University of Twente. Referee:. dr.ir. R.J.M. Westerhof. Suster BV. Members:. prof.dr.ir. G. Brem prof.dr. K. Seshan prof.dr.ir. W. Prins prof.dr. M. Garcia Perez. University of Twente, ET University of Twente, TNW Ghent University Washington State University. v.

(8)

(9) Dedicated to my father.

(10)

(11) ... If I have seen further it is by standing on the shoulders of Giants. - Sir Isaac Newton (5th February 1676).

(12) Summary (xiii) Samenvatting (xvii) साराांश (xxi). C1. General introduction (1). Materials and methods (19) Evaluating quantitative determination of levoglucosan and hydroxyacetaldehyde in bio-oils by gas and liquid chromatography. C2. C3. (39). C4. Fast pyrolysis of cellulose: interaction between chemistry and transport phenomena in the absence and presence of potassium (51). Appendices: A (161), B (201), C (207), D (233) References (259).

(13) Contents. C5. Fast pyrolysis of lignins with different molecular weight: experiments and modelling (91). C6. C7. On the effect of ether linkages in lignin on the product yields and the molecular weight of pyrolysis oils using HSQC NMR (127). Effect of pressure on the fast pyrolysis of acid-leached bagasse and pine wood: experiments and modelling (139). C8. Conclusions and outlook (153). Acknowledgements (287), List of publications (293) About the author (295).

(14)

(15) Summary Depleting fossil energy resources and the changes in climatic conditions have propelled humanity to find alternative sustainable resources to satisfy the increasing demand. Lignocellulosic biomass is one of the carbon-based renewable resources which can be used to produce fuels and chemicals. Fast pyrolysis is one of the promising technologies in which lignocellulosic biomass is decomposed at ∼500 °C and in the absence of oxygen, to produce oil (∼75%), char (∼12%) and gas (∼13%). Pyrolysis oil can be upgraded to fuel or platform chemicals via downstream processing. Literature suggests that the yields of products and their composition obtained during the pyrolysis are affected by the competition between transport phenomena and chemistry. The research described in this thesis aims at advancing the current understanding of the three simultaneously occurring processes, viz. chemical reactions (decomposition, cracking, polymerisation), heat transfer and mass transfer (evaporation, sublimation, random ejection), and their interplay during the fast pyrolysis of cellulose, lignin and lignocellulosic biomass. For that, a dedicated screen-heater reactor was used, which was designed: 1) to minimise non-isothermality (heating rate: ∼5000 °C s-1 ), 2) to control the reaction time inside the reacting particle by the escape rate of compounds from the reaction zone by variation of the pressure, and 3) to minimise reactions outside the reacting particle by minimising hot vapour residence time (∼20 ms) and fast quenching of the products (∼180 °C). Experiments were also carried out in a bench-scale 1 kg h-1 fluidised bed unit equipped with a fractional condensation system to investigate the effect of hot vapour residence time (∼2 s). Qualitative and quantitative characterisation of pyrolysis products was done using various analytical tools. In order to understand the reaction mechanisms, quantitative analysis of species present in pyrolysis oil is necessary. The suitability of gas chromatography (GC) and liquid chromatography (LC) for the quantification of levoglucosan (LG) and hydroxyacetaldehyde (HA) was evaluated. It was found that both GC and LC can principally determine LG quantitatively in pyrolysis oils. However, depolymerisation of oligo-anhydrosugars in GC owns a risk of overestimation of the LG yields. HA can only be determined quantitatively by LC because of its reactions during the high temperature (∼250 °C) injection in GC. It was found that in the absence of potassium salts, cellulose could almost entirely be converted to anhydrosugars while producing hardly any gas (<1%) and char (<1%) in. xiii.

(16) a wide temperature range of 450 to 765 °C. Depolymerisation of cellulose to anhydrosugars was identified to be a true primary reaction; gas and char formation secondary. Mathematical models were developed, including the interaction between chemistry, heat transfer and mass transfer. The escape rate of products from the hot reacting particle was identified as a crucial process affecting the DP distribution of anhydrosugars. The potassium concentration in cellulose was varied to mimic the mineral-rich and pre-treated feedstock and subsequently pyrolysed at 530 °C in screen-heater and fluidised bed. Potassium was found to be catalytically active even when the escape rate of the product away from the reaction front was extremely high (milliseconds). The yields of oil and anhydrosugars decreased significantly as a function of potassium concentration, while the production of other products (water, gases, light oxygenated compounds) was enhanced. The production of char was found to be independent of the escape rate of products at any given potassium concentration. It could be concluded that pyrolysis at reduced pressure, i.e. by fast removal of products from the hot reaction zone, can improve the oil and sugar yields, but only for low alkali and alkaline earth metals content feedstock. The effect of molecular weight on the competing physio-chemical processes was investigated by pyrolysing 14 lignins (350 – 1900 Da) in the screen-heater at 0.5 kPa and 100 kPa. A population balance model was developed, which includes simultaneously occurring cracking reactions, polymerisation reactions and mass transport away from the reaction zone. The model was able to predict all experimentally observed trends after parametrisation. The molecular weight distribution was found to be one of the crucial characteristics of the lignin feedstock, which has a significant influence on the pyrolysis product distribution. Upwards 530 °C, the temperature turned out to have only a minor influence on the yields and composition of the oils produced, whereas the system pressure was identified as the main steering wheel to manipulate the product yields and the molecular weight of the oils. HSQC NMR analysis of lignins and its oils showed that the yields of oil and char and the number average molecular weight of oils were found to be independent of the number of ether linkages in the lignin feedstock. Fast pyrolysis of acid-leached bagasse was carried out in the pressure range of 0.005 to 100 kPa in screen-heater. At the lowest pressure, the total yield of C6 -anhydrosugars (sum of DP1 to DP6 ) was as high as 73% of the poly-C6 -sugars in the feedstock. A mathematical model, including again reaction and mass transfer away from the reaction zone, was able to predict the measured decrease in a total yield of C6 -anhydrosugars and the shift to lighter C6 -anhydrosugars as a function of increasing pressure. At identical pressure and temperature, the total yield of C6 -anhydrosugars obtained from acid-leached xiv.

(17) pinewood was the same for the screen-heater and fluidised bed. As a result of the longer hot vapour residence time in the fluidised bed, DP≥2 C6 -anhydrosugars depolymerised towards DP1 . In nutshell, besides the heating rate of sample, hot vapour residence time of products and the temperature during the pyrolysis, the system pressure is the key parameter, which alters the residence time of products in/on the hot reacting particle, thereby, providing a means to steer the yields and composition of the products of pyrolysis.. xv.

(18)

(19) Samenvatting Het uitputten van fossiele energiebronnen en de klimaatveranderingen hebben de mensheid ertoe aangezet om alternatieve duurzame bronnen te vinden om aan de toenemende vraag te voldoen. Lignocellulose biomassa is een van de hernieuwbare bronnen op basis van koolstof die kunnen worden gebruikt om brandstoffen en chemicaliën te produceren. Snelle pyrolyse is een van de veelbelovende technologieën waarbij lignocellulosische biomassa wordt afgebroken bij ∼500 °C en in afwezigheid van zuurstof, om olie (∼75%), kool (∼12%) en gas (∼13%)te produceren. Pyrolyse-olie kan worden opgewaardeerd tot brandstof of platformchemicaliën via vervolgstappen in het proces. In de literatuur wordt gesuggereerd dat de opbrengsten en de samenstelling van producten verkregen tijdens de pyrolyse, worden beïnvloed door een samenhang vantransportverschijnselen en chemie. Het onderzoek dat in dit proefschrift beschreven wordt, is gericht op het bevorderen van het huidige begrip van de drie gelijktijdig optredende processen. Deze drie processen zijn: chemische reacties (ontleding, kraken en polymerisatie), warmte- en massaoverdracht (verdamping, sublimatie en willekeurige uitstoot) en hun wisselwerking tijdens de snelle pyrolyse van cellulose, lignine en lignocellulose biomassa. Voor het onderzoek werd een speciale screen-heater reactor gebruikt, die ontworpen was om: 1) niet-isotherm gedrag te minimaliseren (verwarmingssnelheid: ∼5000 °C s-1 ), 2) de reactietijd in het reagerende deeltje door de ontsnappingsnelheid van componenten uit de reactiezone te controleren door variatie van de druk, en 3)reacties buiten het reagerende deeltje te minimaliseren door de verblijftijd van hete damp te minimaliseren (∼20 ms) en de producten snel af te koelen (∼-180 °C). Er werden ook experimenten uitgevoerd in een 1 kg h-1 wervelbed opstelling uitgerust met een fractioneel condensatiesysteem om het effect van de verblijftijd (∼2 s) van de hete damp te onderzoeken. Kwalitatieve en kwantitatieve karakterisering van pyrolyseproducten werd uitgevoerd met behulp van verschillende analytische hulpmiddelen. Om de reactiemechanismen te begrijpen, is kwantitatieve analyse van de in pyrolyse-olie aanwezige componenten noodzakelijk. De geschiktheid van gaschromatografie (GC) en vloeistofchromatografie (LC) voor de kwantificering van levoglucosaan (LG) en hydroxyacetaldehyde (HA) werd geëvalueerd. Zowel GC als LC kunnen in principe LG kwantitatief bepalen in pyrolyse-oliën. Echter, depolymerisatie van oligo-anhydrosuikers in GC kan resulteren in een overschatting van de LG opbrengsten. HA kan alleen kwantitatief worden bepaald door middel van LC vanwege reacties tijdens de injectie op hoge temperatuur (∼250 °C) in GC. xvii.

(20) Cellulose kon bij afwezigheid van kaliumzouten vrijwel volledig worden omgezet in anhydrosuikers, terwijl er nauwelijks gas (<1%) en kool (<1%) werd geproduceerd over een breed temperatuurbereik van 450 tot 765 °C. Depolymerisatie van cellulose naar anhydrosuikers werd geïdentificeerd als een primaire reactie; gas- en koolvorming als secondaire reacties. Wiskundige modellen werden ontwikkeld, inclusief de interactie tussen chemie, warmte-t en massaoverdracht. De ontsnappingssnelheid van producten uit het verwarmde reagerende deeltje werd geïdentificeerd als een cruciaal proces dat de DP -verdeling van anhydrosuikers beïnvloedde. De kaliumconcentratie in cellulose werd gevarieerd om de mineraalrijke en voorbehandelde biomassa na te bootsen en vervolgens werd de cellulose bij 530 °C in de screen-heater en wervelbedreactor gepyrolyseerd. Kalium bleek katalytisch actief te zijn, zelfs wanneer de ontsnappingssnelheid van het product van de reactie zone af extreem hoog was (milliseconden). De opbrengsten van olie en anhydrosuikers namen aanzienlijk af als functie van de kaliumconcentratie, terwijl de productie van andere producten (water, gassen en kleine geoxygeneerde verbindingen) was toegenomen. De productie van kool bleek onafhankelijk te zijn van de ontsnappingssnelheid van producten bij elke gegeven kaliumconcentratie. Er zou kunnen worden geconcludeerd dat pyrolyse bij verlaagde druk, d.w.z. door snelle verwijdering van producten uit de hete reactiezone, de olie- en suikeropbrengsten kan verbeteren, maar alleen voor biomassa met een laag gehalte alkali- en aardalkalimetaal. Het effect van het molecuulgewicht op de competitie tussen fysisch-chemische processen werd onderzocht d.m.v. het pyrolyseren van 14 lignines (350 - 1900 Da) in de screen-heater bij 0.5 kPa en 100 kPa. Er is een populatiebalansmodel ontwikkeld dat gelijktijdig optredende kraakreacties, polymerisatiereacties en massatransport uit de reactiezone omvat. Het model was in staat om, na parametrisatie, alle experimenteel waargenomen trends te voorspellen. De molecuulgewichtsverdeling bleek een van de cruciale kenmerken van de lignine-grondstof te zijn, en heeft een significante invloed op de distributie van pyrolyseproducten. Een temperatuur groter dan 530 °C bleek slechts een geringe invloed te hebben op de opbrengsten en samenstelling van de geproduceerde oliën, terwijl de systeemdruk werd geïdentificeerd als de belangrijkste variabele om de productopbrengsten en het molecuulgewicht van de oliën te manipuleren. HSQC NMRanalyse van lignines en bijbehorende oliën toonde aan dat de opbrengsten van olie en kool en de getalgemiddelde molecuulmassa van oliën onafhankelijk zijn van het aantal etherbindingen in de lignine-grondstof. Snelle pyrolyse van zuur-uitgeloogde bagasse werd uitgevoerd in het drukbereik van 0.005 tot 100 kPa in de screen-heater. Bij de laagste druk was de totale opbrengst xviii.

(21) aan C6-anhydrosuikers (som van DP1 tot DP6 ) maximaal 73% van de poly-C6-suikers in de voeding. Een wiskundig model, dat opnieuw reactie en massaoverdracht van de reactiezone af omvat, was in staat om de gemeten afname in een totale opbrengst aan C6-anhydrosuikers en de verschuiving naar lichtere C6-anhydrosuikers als een functie van toenemende druk te voorspellen. Bij identieke druk en temperatuur was de totale opbrengst aan C6-anhydrosuikers, verkregen uit met zuur uitgeloogd dennenhout, hetzelfde voor de screen-heater en de wervelbedreactor. Als gevolg van de langere verblijftijd van hete damp in het wervelbed werden DP≥2 C6 -anhydrosuikers gedepolymeriseerd naar DP1 . Kortom de systeemdruk is de belangrijkste parameter die de verblijftijd van producten in/op het heet reagerende deeltje verandert, waardoor de opbrengsten en samenstelling van de producten van pyrolyse gestuurd kunnen worden. Andere parameters zijn de verwarmingssnelheid van het sample, de verblijftijd van warme dampen van producten en de temperatuur tijdens de pyrolyse.. xix.

(22)

(23) सारांश जीवाश्म उजार् संसाधनांतील सततची होणारी घट आ￱ण हवामानात होणारी स्थत्यंतरे या बाब मुळे मानवाला, उजचे पयार्यी ोत शोधण्याची गरज िनमार्ण झाली आहे. या पाश्वर्भूमीवर लग्नोसेल्युलो￭सक बायोमास (उदा. उसाची ￸चपाडे, वाळलेले गवत) हा एक काबर् न आधा रत उजचा ोत आहे ज्याचा उपयोग इंधन तसेच रसायने तयार करण्यासाठी होऊ शकतो. फास्ट पायारोलीसीस (Fast Pyrolysis) हे एक असामान्य तंत्रज्ञान आहे ज्याद्वारे लग्नोसेल्युलो￭सक बायोमासचे पांतर, ५०० डी. से.ला आ￱ण प्राणवायूच्या अनुप स्थतीत, तेल (७५%), कोळसा (१२%) व इंधन-वायूत (१३%) असे करता येते. पायारोलीसीस तंत्रज्ञाना दरम्यान तयार होणारी उत्पादने आ￱ण त्यांची घटक रचना ही उष्णता हस्तांतरण, वस्तुमान हस्तांतरण व रसायन शा ह्यांच्या मधील स्पधने प्रभािवत होते. प्रस्तुत प्रबंधात केलेले व मांडलेले संशोधन एकाच वेळी होणर्या रासायिनक अ￱भिक्रया, वस्तुमान हस्तांतरण आ￱ण उष्णता हस्तांतरण यावर प्रकाश टाकते. सेल्युलोस, लिग्नन आ￱ण लग्नोसेल्युलो￭सक बायोमास यांच्या पायारोलीसीस मध्ये असलेल्या स्पधचा अभ्यास या प्रबंधात सादर केला आहे. या संशोधनामध्ये स्क्रीन-िहटर रीअक्टरचा वापर केला गेला आहे. त्याची प्रमुख वै￱शष्ट्ये पुढील प्रमाणे आहेत. (१)उष्णता हस्तांतरण दर ५००० डी.से. प्र￸त सेकंद (२) रीअक्टर मधील दाबात बदल क न लग्नोसेल्युलो￭सक बायोमास अंतगर् त व बाह्य अश्या दोन्ही रासायिनक अ￱भिक्रयांना लागणारा वेळ िनयंित्रत करता येतो. (३) रासायिनक अ￱भिक्रयांमधून तयार झालेली उत्पादने द्रािवत नायट्रोजन (-१९६ डी.से.) वाप न थंड केली जातात. सेल्युलोस, लिग्नन आ￱ण लग्नोसेल्युलो￭सक बायोमास चा फास्ट पायारोलीसीस िनःसन्देहपणे एकाच वेळी होणर्या रासायिनक अ￱भिक्रया, वस्तुमान हस्तांतरण आ￱ण उष्णता हस्तांतरण यांच्यातील आं तरिक्रयेद्वारा िनयंित्रत होतो. आधीपासूनच मािहत असल्याप्रमाणे, उच्च उष्णता हस्तांतरण दर तेलाचे उत्पादन वाढिवण्यासाठी आवश्यक आहे. बायोमास बाहेरील घडणार्या रासायिनक अ￱भिक्रयांच्या वेळेचा (१ - २सेकंद) साखर उत्पादानावारती नाममात्र प रणाम घडतो. लिग्ननचे आ ण्वक वजन (िवतरण) हा एक मह वाचा गुणधमर् असून त्याचा पायारोलीसीसच्या उत्पादनांच्या िवतरणावर मह वपूणर् प्रभाव पडतो. पायारोलीसीसचे तापमान ५०० डी. से. पेक्षा कमी असल्यास तेल उत्पादनावर लक्षणीय प्रभाव पडतो. ५५० डी.से. च्यावर पायारोलीसीसचे तापमान गेल्यास रासायिनक अ￱भिक्रयांचा दर हा उष्मा हस्तांतरण दरापेक्षा जास्त होतो ज्या परत्वे पायारोलीसीसचे तापमान न वाढता स्थर रहाते. या सवार्ंव्य￸त रक्त पायारोलीसीस दरम्यान असलेला वायूचा दाब हा मह वाचे प रमाण असून त्याचा उपयोग तेलाचे उत्पादन आ￱ण त्यातील घटकांची संरचना िनयंित्रत करण्यासाठी करता येऊ शकतो.. xxi.

(24) C1.

(25) C1 General introduction. C1.

(26) C1. 1.1. Motivation. The world population is estimated to reach 9.15 billion by the end of 2040, and with that, the energy demand is expected to grow annually at a rate of 1%[1] . Currently, fossil fuels are the primary energy sources, and their extensive use has resulted in the emission of greenhouse gases and particulate matter, causing air pollution and global warming. Uneven distribution of fossil energy resources across the globe and geopolitics dictates the supply and price of energy. Climatic change and energy security, primarily, have stimulated the search for the alternative, CO2 neutral (or negative), energy sources to meet the growing energy demand and to mitigate the carbon footprint. Solar energy, wind energy, hydropower, geothermal and biomass have been identified as renewable resources. Among these, the availability of the first four energy sources is very much dependent on the geographical locations[2,3] . Biomass, being the only renewable carbon source (∼45% C on dry ash free basis[4] ), may fit into the existing fossil infrastructure[5] , and could partially replace crude oil derived fuels and materials. Also, biomass energy is essentially CO2 neutral, if none or little fossil energy is used for its production and processing since it is based on a short carbon cycle[5] . Currently, 12 – 13% of the total world energy demand is satisfied by biomass[5] . However, biomass is often available as low bulk density solids, especially agricultural/forestry waste, and it needs a processing step before it can be transported and used.. 1.2. Lignocellulosic biomass. The 1st generation biofuels (bio-diesel, bio-ethanol), produced from food crops, have been commercialised to mature technologies. In the EU alone, the production of biofuels has increased from 625 (2000) to 35311 kTon y-1 (2017)[6] . However, growing concern about the competition between the increasing demand for food and fuel has shifted the focus towards the 2nd generation biomass feedstock, i.e. lignocellulosic biomass. It can be subdivided into three types namely; softwood (e.g. pine, spruce, cedar), hardwoods (e.g. oak, beech, maple.) and grasses (e.g. miscanthus, switchgrass, reed canary). Lignocellulosic biomass is composed of cellulose, hemicellulose, lignin, extractives and inorganics. They are discussed briefly hereafter since next to biomass their pyrolysis behaviour is investigated in this thesis. Cellulose: Cellulose, an essential structural component of the primary cell wall, accounts for 35 to 50 wt. % of dry lignocellulosic biomass. It is the most abundant, naturally occurring polymer in the biosphere[7] . It is a linear polymer composed of anhydroglucopyranose units connected by β-(1-4) glycosidic bond[8] . The degree of polymerisation 2.

(27) of native cellulose is between 10000 to 14000, whereas it is much lower (200 – 3000) in microcrystalline cellulose (isolated via a pulping process)[7] . Hemicellulose: A second major component of lignocellulosic biomass is hemicellulose. It accounts for ∼28% and ∼35% of dry lignocellulosic biomass in softwoods and hardwoods, respectively[9] . Hemicellulose is composed of three hexoses (glucose, mannose and galactose) and two pentoses (xylose and arabinose), together with some uronic acids[8] , which makes it a hetero-polysaccharide. Softwood hemicelluloses are rich in galacto-glucomannan, and hardwood hemicelluloses are rich in xylans[9] . Unlike cellulose, it is a branched polymer and has a degree of polymerisation in the range 50 to 200[7,9] . Lignin: Lignin is a three dimensional, highly cross-linked, hydrophobic, heteropolymer[10] . It is the most abundant naturally occurring aromatic polymer and accounts for up to ∼35% of dry lignocellulosic biomass. Coumaryl-, coniferyl- and sinapyl alcohols, the precursors of lignin, are interconnected by various linkages, which are broadly classified as ether linkages and carbon linkages[11–13] . During the isolation process (e.g. kraft, organosolv[14,15] ), native lignin undergoes structural changes, and the extent to which the structure of native lignin is preserved is controlled by the severity of the isolation method[15,16] . Note that milled wood lignin, obtained by Björkman’s method, is considered to represent native lignin most accurately[17] . Extractives: In addition to the above three main constituents, extractives are present in the biomass (∼5% on dry mass). They are low molecular weight species, including fats, waxes, alkaloids, proteins, phenolics and essential oils[7,9] . Extractives act as intermediates in plant metabolism, energy reserves, and defences against microbial threats[9] . These can be extracted by using polar (e.g. water, alcohols) or non-polar (e.g. toluene, hexane) solvents[7,9] . Inorganics: The content of inorganic species (also known as ash) in biomass varies significantly as a result of soil and climate conditions[18] . In woody biomass, it is typically in the range between 0.1 to 7.1 wt. %, while in grasses, it can be as high as 42 wt. %[18] . Following are commonly found elements in woody biomass: calcium (Ca), potassium (K), sodium (Na), silicon (Si), magnesium (Mg), aluminium (Al), sulphur (S), iron (Fe), phosphorous (P), manganese (Mn)[18,19] . These minerals are attached to the biomass matrix by an ionic or a covalent bond[19] . The content of alkali and alkaline earth metals (AAEMs), i.e. Na, K, Ca, Mg, in biomass is especially of importance for thermo-chemical conversion process[20] .. 3. C1.

(28) C1. 1.3. Biomass conversion routes. Figure 1.1 shows an overview of different conversion processes to convert lignocellulosic biomass to fuels and chemicals. Direct combustion of biomass to produce energy has been the first step in the history of human civilisation. Worldwide, in large scale plants, direct combustion of biomass is carried out in a fixed bed reactor or grate furnace[21–24] . The main processes and routes depected in Figure 1.1 are briefly discussed here. In gasification, biomass is converted to CO, CO2 , H2 , and CH4 at high temperatures (>900°C) in the presence of air, oxygen or steam. Depending on the relative composition of these gasses, they are referred to as synthesis gas or producer gas[25] . In the liquefaction process, biomass is converted in a temperature range of 150 – 360 °C and at elevated pressures (90 – 250 bar) to yield two main products, liquid (organic + aqueous) and gas[26,27] . The organic fraction is considered as the targeted product of liquefaction and has a significantly lower oxygen content than the feedstock (10 – 25 wt. % in organic phase vs 45 wt. % in feed)[28] . The organic part of the liquid can be upgraded into diesel/gasoline range fuels by catalytic hydrocracking in the presence of a solvent[27] . In the following section, pyrolysis technology, being the focus of this thesis, is elaborated in detail.. 4.

(29) Figure 1.1: Biomass conversion routes for the production of fuels and chemicals [24]. C1. 5.

(30) C1. 1.4. Pyrolysis. The word pyrolysis is based on Greek-derived elements; Pyr (πυρ) implies fire and Lysi (λυση) implies releasing, meaning releasing by fire. In pyrolysis lignocellulosic biomass is converted in the absence of oxygen, to form char, gas, and oil. The yields of these lumped products are affected by the type of feedstock, its ash content, reactor type, heating rate, operating conditions, (hot) vapour residence time of the products[4] . Various pyrolysis modes are available to selectively maximise the yield of the product(s) of interest, i.e. char (charcoal), oil or gas[29,30] . These modes are differentiated based on the size of particles, temperature, heating rate and, most importantly, the vapour residence time of products, as shown in Table 1.1. Table 1.1: Typical product yields (dry wood basis) obtained by different modes of pyrolysis[29,30] Mode Torrefaction Intermediate Fast Gasification a b. Temperature ∼290 °C ∼500 °C ∼500 °C 750 – 900 °C. RT a ∼10 – 60 min ∼20 min ∼10 s seconds. HVRT b – 10 – 30 s 1–2s ∼5 s. Oil 0 – 5% 50% 75% 5%. Char 80% 25% 12% 10%. Gas 15 – 20% 25% 13% 85%. Reaction time Hot vapour residence time. The char, black carbonaceous material rich in minerals, is widely used for the combined heat and power (CHP) generation or as an additive to soil as a fertiliser and carbon sequestration agent[31] . The gas produced during pyrolysis is mainly composed of CO, CO2 and light hydrocarbons, and can be burned to produce process heat. Oil (or biooil or pyrolysis oil) is a complex mixture of organic compounds of different chemical classes such as acids, aldehydes, ketones and anhydrosugars, which are generally present in low concentrations. As an example, typical properties and composition of the fast pyrolysis oil derived from wood are presented in Table 1.2. Compared to crude oil, the composition of pyrolysis oil differs significantly. The relatively high oxygen content in pyrolysis oil makes it thermally unstable, making its direct use as a fuel in CHP systems, burners/engines difficult[32,33] . However, significant efforts have been made to upgrade pyrolysis oil (catalytically) to transportation fuel[34–37] and use as a source for the production of chemicals[38–40] .. 6.

(31) Table 1.2: Typical composition of pinewood derived pyrolysis oil and its properties[32,41] Chemical class Water Acids Alcohols Aldehydes and ketones Sugars LMM lignin HMM lignin Extractives Property pH Specific gravity Elemental composition C H O N Ash HHV Viscosity (at 42°C) Solids Distillation residue. 1.4.1. Unit wt. % wt. % wt. % wt. % wt. % wt. % wt. % wt. % Unit. Value 24 4 2 15 34 13 2 4 Value 2.5 1.2. wt. %. MJ kg-1 cP wt. % wt. %. 54 – 58 5.5 – 7 35 – 40 0 – 0.2 0 – 0.2 16 – 19 25 – 1000 0.2 – 1 up to 50. Pyrolysis reactors. The historical developments in the fields of pyrolysis reactors have been summarised by Garcia-Nunez et al.[42] . Reviews of Butler et al.[43] and Venderbosch et al.[41] have summarised the commercial scale efforts in the fast pyrolysis of biomass. Different fast pyrolysis reactor configurations have been implemented on commercial scale: bubbling fluidised bed reactor (Dynamotive (Canada) 200 tpd, Biomass Engineering Ltd. (UK) 4.8 tpd, Agri-Therm (Canada) 1 – 10 tpd), circulating fluidised bed reactor (Ensyn (Canada) 100 tpd, Metso Consortium (Finland) 7.2 tpd), rotating cone reactor (BTG (Netherlands) 120 tpd, Bio-oil Holding N.V. (Belgium/Netherlands) 12 tpd), auger reactor (KIT-Lurgi (Germany) 12 tpd, ARBI-Tech (Canada) 50 tpd, Renewable Oil International LLC (USA) 4.8 tpd), ablative reactor (PyTec (Germany) 6 tpd). At the time of writing this thesis, a consortium of TechnipFMC, BTG-BTL and Green Fuel Nordic Oy has decided to build four 120 tpd pyrolysis units in Finland[44] . Although fast pyrolysis technology is close to commercialisation, the following are the main chal7. C1.

(32) C1. lenges: 1) scale-up of the technology from laboratory scale concepts/setups to commercial plants, 2) design and operational difficulties, 3) flexibility to accommodate different feedstock (wet/dry, high ash content), and 4) production of quality product to find suitable applications[41,43] . Table 1.3: Characteristics of various laboratory-scale setups used for pyrolysis Reactor – TGA Tubular. Heating rate °C s-1 ∼3 10. Temperature °C 30 – 900 200 – 900. Pressure kPa 100 100. HVRT a – seconds milliseconds. Micro-pyrolyser PHASR c Screen-heater. ∼200 11800 5000. 400 – 600 400 – 500 320 – 800. 100 – 0.005 – 100. seconds ∼10 ms ∼20 ms. Analysis b – Online Online/ offline Online Online Offline. a. Hot vapour residence time of products Possibility of analysing pyrolysis products c Pulse heated analysis of solid reactions b. In laboratories, mainly, thermogravimetric analyser (TGA), tubular reactor, micropyrolyser, PHASR reactor and screen-heater reactor (also called wiremesh) are used: 1) to study the yield of lumped products (gas, char, oil) and their composition of varies types of biomass, 2) to measure (lumped) reaction kinetics, and 3) to investigate ongoing physio-chemical processes and primary/early stage reactions of biomass pyrolysis[45] . The main characteristics of various reactors types used in laboratories are as follows; sample heating rate, operating conditions, (hot) vapour residence time of products and the possibility of analysing formed pyrolysis products, see Table 1.3. For further reading on these reactors, the reader is referred to the review of Sribala et al.[45] . It is important to note that, especially in TGA and micropyrolysers , the actual heating rate of the sample can be significantly different from the one claimed by the machine manufacturer[46–49] . In the screen-heater, nearly isothermal operation is achieved by applying a high heating rate (5000 °C s-1 ). Other important characteristics of the screen-heater are: 1) the escape rate of products, away from the hot reacting particle, can be controlled by manipulating pressure during the pyrolysis, 2) rapid quenching of products is achieved by a liquid nitrogen bath to minimize outside the reacting particle, and 3) importantly, not only light but also heavy species in the oil are collected and can subsequently be analysed. For example, compounds having a molecular weight up to 30000 Da and anhydrosugars with a degree of polymerisation up to 11 can be analysed by using GPC and 8.

(33) LC/MS, respectively. The combination of these characteristics can be achieved using the screen-heater, making it suitable for studying primary reaction products.. 1.4.2. Chemical and physical processes during pyrolysis. Figure 1.2: Schematic representation of chemical and physical processes occurring during pyrolysis Lédé[47] suggested that the pyrolysis of biomass and its building blocks is governed by the competition between physical and chemical processes. During pyrolysis, a fraction of lignocellulosic biomass passes through a liquid intermediate (Figure 1.2), also called a melt, which has a lower degree of polymerisation compared to that of the original polymers. In the case of cellulose, this intermediate liquid is termed the active cellulose[50] . At the reacting particle and/or within this liquid intermediate[51] , a large number of chemical reactions take place, which are broadly classified into cracking reactions to form volatiles and gases, and polymerisation reactions to produce char. Volatile products are transported away from the hot reaction front not only by evaporation but also by random ejection in the form of aerosols[52–55] or by sublimation[56] . The mass transport rate of products depends on two factors, i.e. the molecular weight of species and the operating conditions during the pyrolysis. At a constant pyrolysis temperature, smaller molecules have higher vapour pressure due to which they can escape the hot reaction front easily. On the contrary, because of the lower mass transport rates, bigger molecules spend a longer time at the hot reaction front, where they consequently undergo cracking as well as polymerisation reactions. Once escaped from the hot reaction front, products may undergo secondary reactions in/on the surface of 9. C1.

(34) C1. nascent char (containing AAEMs) or in the hot vapour phase before condensation. Secondary reactions of the vapours facilitate the formation of gasses and light oxygenated compounds. The extent of secondary reactions depends on the hot vapour residence time[57,58] , temperature[58] and the presence of AAEMs[58] . Next to the mass transport of species, the supply of energy to drive the pyrolysis reactions is of equal importance. The heat transfer is strongly influenced by the size of the biomass particle, surrounding temperature and the transfer mechanism[59] . Pyle et al. have identified the following three regimes, based on Biot number (Bi) and Pyrolysis (Py1 and Py2 ) numbers[60] , in which pyrolysis reactions are controlled by: 1) kinetics (reactions taking place at the surrounding temperature), 2) External heat transfer (absence of heat carrier from the heat source to the particle), and 3) Internal heat transfer (big particles). In essence, the heating rate of the lignocellulosic biomass predominantly affects the set of pyrolysis reactions it will undergo, resulting in a different outcome. Collectively, the yields and the distribution of pyrolysis products are influenced by the interplay between reaction rates, the heating rate of the particle and the rate of mass transport of species/products away from the hot reaction front.. 1.4.3. Role of AAEMs during pyrolysis. The AAEMs, present in the biomass matrix, reportedly affect the carbohydrate fraction (cellulose and hemicellulose) of the biomass by catalysing ring-fragmentation and dehydration reactions[61,62] . As a result of that, the yields of anhydrosugars and organics decrease while facilitating the production of light oxygenated compounds, water and gaseous species (e.g. CO, CO2 , CH4 ). The yields of organics and anhydrosugars can be improved by: 1) minimising AAEMs content by pre-treating the lignocellulosic biomass with (hot) water[63] or by acid-water mixture (mineral acid[61,64,65] or organic acid[62,66,67] ), and 2) by ceasing the catalytic activity of AAEMs by passivating them by strong acids (H2 SO4 , H3 PO4 ) to form stable salts which are catalytically less active[68–70] . However, it is unknown to what extent the destructive effects of AAEMs on sugar production can be minimised by maximising the product removal rate. In the literature, Oudenhoven et al.[71] have reported the maximum sugar yields of 30% on acid-leached pinewood basis in a fluidised bed, whereas others have reported the sugar yields in the range between 8% and 22%[61,64,72–75] . Pyrolytic sugars can be upgraded to monosaccharides (e.g. glucose, xylose, mannose) via hydrolysis followed by fermentation and catalytic processing to produce bio-ethanol and platform chemicals[76] , respectively.. 10.

(35) 1.4.4. C1. Analytical techniques. A combined use of qualitative and quantitative analysis of pyrolysis oil is needed: 1) in understanding its composition, 2) for proposing reaction mechanisms, 3) for optimising process conditions to maximise the production of the targeted product(s), and 4) for determining its physical properties for downstream applications such as catalytic upgrading. Reviews[77–79] available in the literature, provide a detailed overview of analytical techniques and strategies used for the characterisation of pyrolysis oils. Gas chromatography (GC) with mass spectrometry (MS) is used to analyse the volatile fraction (Tboiling ∼320 °C) of pyrolysis oil. In principle, whole pyrolysis oil, i.e. water-soluble (e.g. sugars, acids, aldehydes) and water-insoluble (e.g. mono-, oligo-phenolics), can be characterised by liquid chromatography (LC) with MS. Gel permeation chromatography (GPC) or size exclusion chromatography (SEC) provides a molecular weight distribution of pyrolysis oil. Additionally, different spectroscopic (e.g. NMR, ESR) and spectrometric (e.g. mass, UV-vis, FTIR) techniques are used extensively to characterise pyrolysis oil.. 11.

(36) C1. 1.4.5. Modelling. Figure 1.3: Multi-scale modelling for the pyrolysis of biomass Mathematical modelling proves to be a useful tool for process design, development and optimisation, and predicting reactor performance and product yields[4] . In a few extensive reviews[4,59,80–82] , efforts of biomass pyrolysis modelling have been summarised. Pyrolysis of biomass is a multi-scale modelling problem (Figure 1.3), encompassing reactions at the molecular level to the hydrodynamics in the reactor. One/multi component (lumped) kinetic models are based on the weight loss profiles of biomass (or its building blocks) measured under slow pyrolysis conditions[83–91] . In transport models, the physical processes (mass, heat and momentum transfer) are taken into account across the biomass particle where volume and the physical properties vary with the degree of conversion and temperature[92–100] . Next, reactor models, a combination of previous two models and hydrodynamics, are available in the literature for fixed bed reactor[101,102] , rotating cone reactor[103] , fluidised bed reactor[104,105] . More recently, density functional 12.

(37) theory (DFT) based calculation methods have been exploited to understand and describe the pyrolysis reaction mechanisms of cellulose[106–108] , hemicellulose[109,110] and lignin pyrolysis[111–113] . Independently developed lumped reaction kinetics models and transport models, when combined, can be used to predict the trends of the yields of oil, char and gas[4] . However, they possess limited prediction power due to the wide variation in the reported kinetics[114] . Also, these combined models, except those of Solomon et al.[115,116] , do not take into account the effect of molecular weight on the mass transport rate of species. The macroscopic structures of microcrystalline cellulose[117] and lignin[118] are significantly different from that of lignocellulosic biomass[7] , hence, restricting the applicability of transport models to biomass. Most importantly, as described in Section 1.4.2 of this introduction, the mass transport rate dependent likelihood of chemical reactions occurring at the hot reaction front, is not included in the combined lumped reaction kinetics-transport models. Therefore, in the author’s point of view, a unified mechanistic model (yellow shaded area in Figure 1.3) incorporating a simultaneous competition/interaction between chemistry (cracking and polymerisation reactions), heat transfer (sample heating rate) and mass transport of products (by sublimation, evaporation and random ejection) away from the hot reaction front is needed to describe the pyrolysis of biomass and/or its building blocks more accurately.. 13. C1.

(38) C1. 1.5. Outline of the thesis. As mentioned in preceding sections, the yields of products and their distribution obtained during pyrolysis are affected by the relative rates of reactions (cracking and polymerisation), heat transfer and mass transfer. The work presented in this thesis aims at advancing the current understanding of these chemical and physical processes and their interplay during fast pyrolysis of biomass, cellulose and lignin. For that, a significant part of the experimental work was done using a previously developed[119] and characterised[120] screen-heater. It is designed to obtain insights into primary (or early stage) reactions of pyrolysis by minimising non-isothermality by employing high heating rates (∼5000 °C s-1 ). In addition, mass transfer limitations of products were minimised (hot vapour residence time ∼20 ms) by reducing pressure during pyrolysis in combination with rapid quenching of products (∼-180 °C), thereby, minimising secondary reactions outside the reacting particle. The work presented in the subsequent chapters is as follows. Chapter 2. A detailed overview of materials, experimental procedures, experimental setups, and analytical techniques that are used in this thesis is provided in this chapter. Product definitions, yield calculations and mass balances are also described. Chapter 3. This chapter evaluates the pitfalls of using gas chromatography and liquid chromatography in the quantification of levoglucosan (LG) and hydroxyacetaldehyde (HA) in biomass and cellulose derived pyrolysis oils. For that, pyrolysis oil samples infused with known amounts of LG and HA were analysed using both techniques. Chapter 4. This chapter deals with fast pyrolysis of microcrystalline cellulose (Avicel PH 101) in the dedicated screen-heater to study the primary pyrolysis reactions and products. Under these conditions, the effect of temperature on the product yields and the distribution of anhydrosugars in oil is studied. Two mathematical models are developed to interpret and predict experimentally observed trends and to investigate if the pyrolysis is kinetically controlled or if there is an interplay between heat, mass transfer and chemistry. Based on the experimental observations and modelling results, the most prevailing parameters influencing the DP distribution of anhydrosugars are identified. It is widely accepted that the AAEMs have a destructive effect on the production of anhydrosugars. However, it is unknown to what extent this effect can be minimised by maximising the product removal rate from the reacting particle. For that, potassium infused cellulose (concentration: 1 to 10000 mg kg-1 ) samples were pyrolysed, and the effect of potassium on the yields of oil, char, gas and sugars is investigated. The conditions applied in the screen-heater (0.5 kPa and ∼20 ms hot vapour residence time) 14.

(39) are challenging to replicate in an industrial scale reactor. Hence, results obtained from screen-heater are compared with a bench scale fluidised bed. Chapter 5. In this chapter, the interplay of chemistry and mass transfer will be studied using 14 lignins having number average molecular weights between 350 Da and 1900 Da. These lignins were characterised and subsequently pyrolysed in a wide range of operating conditions (Temperature: 425 to 793 °C; Pressure: 0.5 kPa and 100 kPa). Experimentally obtained oil and char yields and the molecular weight (distribution) of oils are interpreted as a function of molecular weight of the feedstock. A population balance model, including three concurrently occurring processes, viz. cracking reactions, polymerisation reactions and mass transport (molecular weight dependent) from the hot reacting particle, was developed. Based on the experimental observations and the modelling results, the most crucial process parameter to steer the yields of oil and char, and the molecular weight of the oil was identified. Chapter 6. Next to the molecular weight of the feedstock lignin, its chemical structure may influence the pyrolysis outcome as well. Five lignins, isolated via three different methods, and their oils (Temperature: 530 °C; Pressure: 0.5 kPa and 100 kPa) are characterised by HSQC NMR to elucidate structural changes. Three main ether linkages, namely, β-aryl ether (β-O-4), phenylcoumaran (β-5’) and resinol (β-β’), were detected and quantified. In this appendix, the influence of these ether linkages on the yields of oil and char, and the molecular weight of products is studied. Chapter 7. Finally, fast pyrolysis of acid-leached biomass is studied by varying the system pressure (0.005 to 100 kPa) in order to investigate the relative effects between the reaction rate and the escape rate of products. The total yield of anhydrosugars and their DP distribution is obtained as a function of pressure. A mathematical model, including the competition between (depolymerisation and decomposition) reactions and mass transport of anhydrosugars, is used to interpret and predict experimentally observed trends. Additionally, the effect of hot vapour residence time on the DP distribution of anhydrosugars is investigated by pyrolysing acid-leached biomass in a fluidised bed reactor. Additionally, a mathematical model is developed to interpret experimentally observed influence of hot vapour residence time on the DP distribution of anhydrosugars. Chapter 8. The main conclusions of this thesis with the outlook are presented in this chapter.. 15. C1.

(40) C1. Appendix A. Supporting information for chapter 3. Appendix B. Supporting information for chapter 4. Appendix C. Supporting information for chapter 5. Appendix D. Supporting information for chapter 7.. 16.

(41) C1. Nomenclature Bi. Biot number. External heat transf er rate hR = λ Internal heat transf er rate. Py1. Pyrolysis number 1. λ Internal heat transf er rate = kρcp R2 Reaction rate. Py2. Pyrolysis number 2. External heat transf er rate h = kρcp R Reaction rate. 17.

(42) C2. 18.

(43) C2 Materials and methods. C2.

(44) Abstract. C2. This chapter provides an overview of the materials, experimental procedures and setups, and analytical techniques used in this thesis. Short, but adequate, descriptions are given and for details the reader is directed to relevant publications.. 20.

(45) 2.1. Chemicals and Materials. All chemicals and materials used in the experiments performed in this thesis are listed in Table 2.1 and Table 2.2, respectively. Table 2.1: Chemicals used in the experiment Compound Acetol (1-Hydroxy-2-propanone) Barium carbonate (BaCO3 ) Cellobiosan (1,6-Anhydro-β-cellobiose) Cellotriosan (1,6-Anhydro-β-D-cellotriose) Cellotetrasan (1,6-Anhydro-β-D-cellotetrose) Dichloromethane (DCM) Glucose Hydroxyacetaldehyde (as glycolaldehyde dimer) Levoglucosan (1,6-Anhydro-β-D-glucopyranose) Levoglucosenone (1,6-anhydro-3,4-dideoxyhex-3-enopyran-2-ulose) Potassium carbonate (K2 CO3 ) Potassium chloride (KCl) Potassium hydroxide (KOH) Sulfuric acid (H2 SO4 ). 21. Purity % >89 >99 >95 >98 >98 ≥99.8 >99.99 >99 >98 >95. Supplier Sigma Aldrich Sigma-Aldrich Carbosynth Ltd. LC Scientific LC Scientific Merck Sigma-Aldrich Sigma-Aldrich Carbosynth Ltd. Carbosynth Ltd.. >99 >99 >98 >99.99. Sigma-Aldrich Sigma-Aldrich Sigma-Aldrich Sigma-Aldrich. C2.

(46) 22. Wheat straw organosolv lignin (WSL). Spruce organosolv lignin (SL). Energy Research Centre of the Netherlands (ECN) Energy Research Centre of the Netherlands (ECN). Shell Prepared in house Rettenmaier & Söhne GmbH Prepared in house Sigma-Aldrich. Bagasse Milled wood lignin (MWL) Pinewood (lignocel 9) Pyrolytic lignin (PL) Sigma organosolv lignin (SOL). Silica sand. Supplier Sigma-Aldrich. Materials Avicel PH 101 cellulose. Isolated by water-ethanol organosolv process[16]. Comment ∼50 µm, 60.5% crystallinity[121] , 5 mg kg-1 ash, 1 mg kg-1 AAEMs, average DP 220[122] Particle size: <150 µm Isolated by Björkman’s method[15] Particle size: 1) 0.5 to 2 mm, 2) <150 µm Isolated by cold water precipitation (Section 2.2.5) Prepared from a mixture of three hardwoods (50% Mapel, 35% Birch, 15% Poplar) Particle size: 212 to 300 µm, particle density: 2600 kg m-3 Isolated by water-ethanol organosolv process[16]. Table 2.2: Materials used in the experiment. C2.

(47) 23. Compound Acetone Acetonitrile Ammonium acetate (NH4 CH3 CO2 ) Deionised water Deuterated methanol (CD3 OD) Deuterium oxide (D2 O) Dimethyl sulfoxide (DMSO) Methanol Tetrahydrofuran (THF) ≥99.98 >99.9 99.9 atom 99.9 atom ≥99.9 >99.9 ≥99.8. Purity % >99.9. Cambridge Isotope Laboratories Inc. Sigma Aldrich Sigma Aldrich Merck Merck. Supplier Merck Merck Sigma Aldrich. Table 2.3: Chemicals used in the characterisation. The chemicals used for the characterisation are listed in Table 2.3.. Analysis GC-FID/MS HPLC Direct infusion MS HPLC 1H NMR 1H NMR Q-CAHSQC NMR GC-FID/MS GPC. C2.

(48) C2. 2.2. Methods. 2.2.1. Impregnation of salts in cellulose and levoglucosan. K2 CO3 , KCl or KOH was dissolved in deionised water, and then dry cellulose was added to the mixture. The potassium (source: K2 CO3 ) concentration in cellulose was varied between 100 mg kg-1 and 10000 mg kg-1 . For KCl and KOH, the concentration of potassium was 1000 mg kg-1 . The water-salt-cellulose mixtures were thoroughly mixed (T = 20 °C) in a round bottom flasks for 1 h. After mixing, deionised water was removed using a rotary evaporator (BÜCHI Rotavapor R-200, T = 65 °C, P = 10 kPa, ∼1 h), and wet cellulose samples were dried for 24 h using a vacuum oven (Heraeus FVT420, T = 20 °C, P = 0.5 kPa). Potassium (source: K2 CO3 ) was impregnated in levoglucosan to have a concentration of 1000 mg kg-1 . Unlike cellulose, levoglucosan readily dissolved in a water-K2 CO3 mixture. Deionised water was slowly removed using the same method as for cellulose, while the mixture is continuously stirred to ensure uniform distribution of potassium. Gradual precipitation of sugar and salt, from the saturated aqueous solution, ensured a homogeneous distribution of potassium in levoglucosan.. 2.2.2. Acid Leaching. The acid leaching of biomass samples was carried out in a 1 L round bottom glass reactor. For acid leaching of biomass, a synthetic mixture mimicking 2nd condenser liquid (aqueous acidic solution) from the pyrolysis process was used, of which the composition is presented in Table 2.4. For details, readers are kindly referred to the literature[62,66,71] . The glass reactor, filled with the leaching solution, was preheated to 90 °C. A watercooled condenser was used to condense the volatile compounds. Biomass was added to the reactor, once the desired temperature was reached, in a mass ratio of 0.1 g of dry biomass to 1 g of the leaching solution. The mixture was thoroughly mixed by a magnetic stirrer (Magneetroerder X-1250) at 600 rpm for 2 h. After that, biomass slurry was vacuum filtered through a 6 µm filter (Whatman quantitative filter paper, ashless, grade 42). The biomass was rinsed with the deionised water, to remove the leaching solution from the pores of the particle, until the pH of the effluent liquid was higher than 6. Acid pre-treated biomass was dried in a vacuum oven (Heraeus FVT420, P <0.5 kPa) at room temperature for 24 h.. 24.

(49) Table 2.4: Chemical composition of synthetic mixture mimicking 2nd condenser liquid Component Acetic acid Ethanol Acetone Propionic acid Guaiacol Deionised water. 2.2.3. Mass percentage % 10 3.75 3.75 1.5 1.5 79.5. C2. Ash content determination. The ash content of a sample was determined by using a dry oxidation method[123] . A crucible was rinsed with deionised water followed by dilute nitric acid (0.1 M). The crucible was dried at 105 °C and stored in a desiccator at room temperature. The weight of the crucible was measured three times with and without sample on a weighing balance (Mettler Toledo AX205, max 220 g, precision 0.01 mg) and the average weight was recorded. Next, the crucible containing the sample was kept in a muffle furnace at 575 + − 5 °C for 24 h. After that, it was cooled down to room temperature in a desiccator and weighted again using the same balance. The ash content of the sample was determined by subtracting the weight of the empty crucible from the crucible including ash, and the weight difference is divided by the sample weight (Eq. 2.1). YAsh =. 2.2.4. Mcrucible+ash − Mcrucible × 100 Mcrucible+sample − Mcrucible. (Eq. 2.1). Acid hydrolysis. The total amount of anhydrosugars present in pyrolysis oils were determined by acid hydrolysis. For that, the NREL LAP method “Determination of sugars, by-products, and degradation products in liquid fraction process samples” was used[124] . Approximately 40 mg of pyrolysis oil was dissolved in 10 g of deionised water. 350 µL of concentrated H2 SO4 (72 wt.%) was added to the diluted oil sample. The mixture was heated in an autoclave at 120 °C for 1 h and was allowed to reach the room temperature. The mixture was transferred to a centrifuge tube, and BaCO3 was added to the mixture to neutralise H2 SO4 . Precipitated barium sulphate (BaSO4 ) was separated by centrifuge (Eppendorf centrifuge 5804, 9000 rpm, 5 min).. 25.

(50) 2.2.5. C2. Cold Water Precipitation. Pyrolytic lignin was produced by cold water precipitation as described in the literature[125,126] . Pinewood-derived pyrolysis oils were added dropwise to ice-cold water while being stirred continuously. The precipitated powder was recovered by vacuum filtration followed by drying in a vacuum oven (Heraeus FVT420, P <0.5 kPa) at room temperature for 24 h. Dry pyrolytic lignin was stored in a freezer maintained at -25 °C.. 2.2.6. Lignin fractionation. Lignins were fractionated into low and high molecular weight fractions by solvent extraction[127,128] . A lignin sample was added to DCM in a weight ratio of 1 to 10. The DCM insoluble fraction, referred to as the heavy fraction (H), was separated from the soluble fraction by filtration. The DCM soluble fraction, referred to as the light fraction (L), was recovered by evaporating DCM. See Figure 2.1 for the schematics.. Figure 2.1: Lignin fractionation procedure. 26.

(51) 2.3. Fast pyrolysis setups. 2.3.1. Screen-heater. Fast pyrolysis of cellulose, lignin and biomass was carried out in the screen-heater setup. The schematic representation of screen-heater is shown in Figure 2.2. The screenheater setup and its main characteristics, sample preparation, and the operating procedure is described here. 1 2 3 4 5 6 7 8 9 10 11 12 13 14. Vessel (glass) Mesh/Sample Pyrometer spot Liquid nitrogen bath (glas) Copper electrode/clamp Tape Thermocouple connection Syringe connection (gas sample) Pressure sensor Vacuum pump Nitrogen gas Pyrometer Tube (glass) Silicon sealing. 2X12 Volt DC High ampere. 7. 9 P. 8. 5 6. 3. 6. 11. 2. 5 1 4. 10. 2. 14. 12. 13. 1 4. Figure 2.2: Schematics of the screen-heater setup. Sample preparation and setup A uniformly distributed sample (∼50 mg) was pressed between the two catalytically inactive[120] screens (#2, thickness: ∼45 µm[119] ) using a 16-tonne hydraulic press. The pressed screens were clamped between two copper electrodes with bolts (#5). The electrodes, clamps and bolts (#5) were covered with tape (#6). A glass vessel (#1, Duran 250 mL centrifuge tube, round bottom, 147 mm x 56 mm) was installed around the copper electrodes (#5) and screens (#2) as shown in Figure 2.2. The reactor with the glass 27. C2.

(52) vessel was placed in the liquid nitrogen (-196 °C) bath (#4) to achieve quick quenching of the produced vapours/aerosols.. C2. The glass vessel was purged with the dry nitrogen gas (#11) to remove air from the vessel before each experiment. Vacuum (≤0.5 kPa) was achieved by an inline vacuum pump (#10). In case of atmospheric pressure experiments (referred to as 100 kPa experiment), a nitrogen atmosphere of ∼100 kPa was maintained in the glass vessel after flushing the air out. The pressure inside the glass vessel was 0.5 kPa or 100 kPa and was monitored by two high-speed pressure sensors (#9, Heise DXD3765 and Druck PTX 520). The vapour phase temperature was recorded by using a K-type thermocouple (diameter 0.025 mm connected to a Weidmuller MAS K-type thermocouple signal conditioner, recording every 16 ms), introduced from #7, placed at 15 mm distance from the screens. The temperature of screens was measured with a pyrometer (#12, Kleiber KGA 730). The detailed calibration procedure of the pyrometer is described elsewhere[120] . The optical beam of pyrometer was focused at the centre of the screens (#3) via a glass tube (#13) and silicon sealing (#14) to avoid the disturbance of the temperature measurement by the liquid nitrogen. The temperature measured at the pyrometer spot was termed as the final temperature of the screens (TFS ), and to reach TFS the screens were heated at a rate of 5000 °C s-1 approaching as close as possible to isothermal measurements[119,120] . The temperature of the screens was measured and regulated by a calibrated pyrometer[120] and a PID control routine programmed in LabVIEW, respectively. The holding time of the screens at a given TFS was set at 5 s unless otherwise mentioned, and after that, they were cooled at a rate of ∼60 °C s-1 . Experiments were performed at 0.5 kPa and 100 kPa unless otherwise mentioned. It was estimated that for both pressures the quenching rate of the formed products was very fast (hot vapour residence time in order of 20 ms), which ruled out a significant effect of reactions outside the reacting sample (minimised vapour phase reactions)[120] . Product recovery and mass balance determination At the end of each experiment, the reactor was allowed to reach room temperature. In reduced pressure experiments, the reactor was filled with dry nitrogen gas before taking a gas sample from the glass vessel (#8), and for 100 kPa experiments, a sample was taken from a gas bag. The composition of gases, determined by using gas chromatography, was used to calculate the gas yield (kg gas kg-1 feedstock), as given by Eq. 2.2. The char yield (kg char kg-1 feedstock) was estimated from the difference between the weights of the screens before (with feedstock), and after (with char) the experiment (Eq. 28.

(53) 2.3).. YGas. Psample fv,i (Vvessel + Vbag )Mw,i RTambient = Mf eedstock (1 − fw − fash ). YChar =. Mscreens+char − Mscreens Mf eedstock (1 − fw − fash ). (Eq. 2.2). (Eq. 2.3). Oil was condensed primarily on the interiors of the glass vessel and to a minor extent on the electrodes, clamps and bolts. The weights of the tape, wrapped around the electrodes, clamps and bolts, and the glass vessel were recorded before and after the experiment and used to calculate the oil yield (kg oil kg-1 feedstock), see Eq. 2.4. Independently determined oil, char, and gas yields were summed together to determine the overall mass balance closure of the experiment (Eq. 2.5).. YOil =. 2.3.2. (Mvessel+oil − Mvessel ) + (Mc+b+t+oil − Mc+b+t ) Mf eedstock (1 − fw − fash ). (Eq. 2.4). M ass balance = YGas + YChar + YOil. (Eq. 2.5). Fluidised bed. Bench-scale fluidised bed Cellulose samples (pure, 100 mg kg-1 and 1000 mg kg-1 potassium impregnated) were pyrolysed in a bench scale fluidised bed reactor at a temperature of 530 °C. Silica sand was used as bed material and preheated nitrogen as the fluidisation gas. During each experiment, ∼100 g of cellulose was fed manually to the reactor in batches of 2 to 5 g together with 4 to 8 g of sand. The feeding system consisted of two valves which functioned as a gas lock. The experimental time was 25 min. Produced char was separated from the outgoing gas/vapour stream leaving the reactor by using a wire-mesh filter (opening size 9 µm). Almost all char, including potassium, was collected on the filter. This filter cake, with varying amounts of potassium, was used as a tool to study the secondary vapour phase reactions. The estimated vapour residence time of products in the hot zone of the setup, including reactor and tubing, was always kept around 1.6 s. About 95% of the produced vapours were condensed by an electrostatic precipitator 29. C2.

(54) C2. (ESP) operated at 20 °C (outgoing gas temperature). The remaining vapours were condensed in a double-walled glass condenser operated at -5 °C (outgoing gas temperature). Both fractions were considered for the mass balance. The char yield was determined by weighing the char/sand mixture and char/filter after the experiment and subtracting the initial weight of the sand and filter plus the sand fed during the experiment. The gas yield was determined by difference. Pilot plant - continuous fluidised bed Fast pyrolysis of (untreated and acid-leached) pinewood, under atmospheric and reduced pressure, was carried out in the continuous fluidised bed, with a capacity of 1 kg h-1 biomass. The schematic representation of the setup is shown in Figure 2.3. All experiments were carried out in a nitrogen atmosphere. The pilot plant included two hoppers each with 4 kg capacity, of which one is used for storing pinewood (#1) and second for the sand (#5). The feeding rates of pinewood and sand screws were calibrated and controlled by two different systems. A mixture of pinewood and sand was fed to the fluidised bed reactor by using a third screw (#3). The temperatures of two hot zones, i.e. fluidised bed reactor (#6) and char separation section, were controlled independently. In all experiments, the temperature of the fluidised bed reactor was maintained at ∼485 °C. The operational conditions are summarised in Table 2.5. An overflow vessel (#7), connected at the bottom of the reactor, was used to collect sand and char particles from the reaction zone. Solid particles, entrained with gas/vapours, were removed by a knocked out vessel (#8) and cyclones (#9). The outlet of cyclones was connected to the jacketed electrostatic precipitator (ESP, #10), where the temperature of the ESP condenser was maintained at 20 °C (outgoing gas temperature). The products, which could not be condensed in ESP, were sent to a second condenser, Table 2.5: Summary of the experimental conditions of reactor and condenser Parameter Run time Msand,initial U/Umf TESP TIC PReactor τ vapors. Unit min kg °C °C kPa s. Normal operation. ∼2. 100 1.9. 30. Reduced pressure operation 90 ∼2 ∼4 20 -10 55 1.

(55) 31. 4. N2. 6. 7. 8. 9. Oven. Char. 10. Bio-oil. Cooling in. Cooling out. 11. Figure 2.3: Schematic representation of pilot plant – continuous fluidised bed. 3. 5. TI. 12. 13. 1) Biomass hopper, 2) Stirrer, 3) Feeding system, 4) Cooling jacket of feeding screw, 5) Sand storage/feeding system, 6) Fluidized bed reactor, 7) Overflow, 8) Knock-out vessel, 9) Cyclones, 10) Electrostatic precipitator (ESP), 11) Intensive cooler (IC), 12) Gas filter, 13) Vacuum pump. 2. 1. N2. Gas. C2.

(56) C2. also called as an intensive cooler (IC, #11), operated at -10 °C. The permanent gases passed through a gas filter (#12) to collect the remaining liquid. The vacuum in the pilot plant was achieved by using a vacuum pump (#13) connected at the outlet of the gas filter (#12). A needle valve was used at the inlet of the vacuum pump to control the pressure in the pilot plant. The volumetric flow rate of the outgoing gas was measured by using a dry gas meter. At every 10 min, a gas sample was taken from the outgoing gas stream. Mass balance determination After each experiment, oil, char and gas yields were individually determined as follows. The char present in the reactor, the overflow vessel, the knock out vessel and the cyclones was collected to determine the char yield (Eq. 2.6). Based on the gas composition and the flowrate of nitrogen gas fed to the fluidised bed, the gas yield (Eq. 2.7) was calculated. The mass of oil collected in ESP and IC were added together to determine the total yield of (wet) pyrolysis oil (Eq. 2.8). The yields of oil, char and gas were summed together to determine the overall mass balance of the experiment (Eq. 2.5).. YChar =. MC,R + MC,O + MC,KD + MC,C Mbiomass (1 − fw − fash ). Z t=tend ( φout Psample. YGas =. t=0. RTambient. φin Psample − RTambient. ) n X. fv,i Mw,i dt. i=1. Mbiomass (1 − fw − fash ). YOil =. MO,ESP + MO,IC Mbiomass (1 − fw − fash ). 32. (Eq. 2.6). (Eq. 2.7). (Eq. 2.8).

(57) 2.4. Analytical techniques. 2.4.1. Gas Chromatography (GC). C2. Micro-GC A gas chromatograph (Varian MicroGC CP4900, 2 analytical columns: 10 m Molsieve 5A, 10 m PPQ, carrier gas: He) was used for analysing gas samples of screen-heater experiments for N2 , O2 , CO, CO2 , CH4 , C2 H6 , C2 H4 , C3 H6 and C3 H8 . The samples were analysed twice to ensure the reproducibility. Refinery GC The gas composition of the samples, collected during the pilot plant experiments, was determined using gas chromatography (Varian 450-RGA). The GC was equipped with three channels. Channel 1 equipped with two columns (Hayesep Q - Agilent CP1305 and Molsieve 5A - Agilent CP1306) detected H2 using thermal conductivity detector (TCD) with N2 as a carrier gas. N2 , O2 , CO and CO2 were separated by three columns (Hayesep Q - Varian CP1308, Hayesep N - Varian CP1307, Molsieve 13X - Varian CP1309), and were detected by channel 2 using TCD with He as a carrier gas. In channel 3, equipped with two columns (CP-Sil 5CB - Varian CP1310 and Al2 O3 /MAPD - Varian CP7433), hydrocarbons (CH4 , C2 H6 , C2 H4 , C3 H6 , C3 H8 ) were detected using a flame ionisation detector (FID) with He as a carrier gas.. 2.4.2. Gas Chromatography-Mass Spectrometry. Pyrolysis oils were diluted with a solvent in a mass ratio of 1:20 and were filtered with a 0.45 µm Whatman RC Agilent filter, and subsequently analysed by using GCFID/MS (GC - 7890A, MS - 5975C Agilent Technologies system) equipped with a Agilent HP-5Ms HP19091S-433 capillary column (60 m, ID 0.25 mm, Film thickness: 0.25 µm). The column is packed with (5%-Phenyl)-methylpolysiloxane. Helium was used as carrier gas with a constant flow of 1.95 mL min-1 . The oven temperature was programmed from 45 °C (4 min) to 280 °C (20 min) at a heating rate of 3 °C min-1 . The injector and the column to MS interface were maintained at a constant temperature of 250 °C and 280 °C, respectively. A sample of 1 µL was injected into the GC. The deactivated inlet liner (Agilent part no: 5183-4711) was regularly replaced to eliminate the influence of non-volatile (inorganic) species deposited on the glass wool. The MS was operated in electron ionisation mode, and ions were scanned in a m/z range from 15 to 500. The identification of the peaks was made by matching its mass spectra with the NIST and Wood library or on the retention times of standards of known compounds 33.

(58) C2. injected in the column. The limit of detection for a given species was 0.008 wt. %, while 0.01 wt. % was the limit of quantification. The quantification was performed based on the FID.. 2.4.3. Ion chromatography. The content of AAEMs (Na+ , K+ , Ca2+ , Mg2+ ) present in a sample was determined by using ion chromatography (IC). The water-soluble fraction of the sample was dissolved in 8 g of deionised water. The samples were filtered (Whatman RC Agilent 0.2 µm filter) before the IC analysis. The IC (Metrosep 850 Professional IC) was equipped with a cation column (Metrosep C6 - 150/4.0) and Metrohm 732 IC detector coupled to an IC Separation Centre. The analysis was performed at room temperature and with eluent (1.7 mM nitric acid + 1.7 mM dipicolinic acid) flow rate of 1 mL min-1 . The quantification was done based on four point calibration of each cation.. 2.4.4. High-pressure liquid chromatography. Liquid phase analysis of pyrolysis oils was performed by using HPLC (Agilent 1200 series). The system was equipped with Hi-Plex Pb column (7.7 x 300 mm, 8 µm), and refractive index detector (RID, T = 55 °C) and variable wavelength detector (VWD, λ = 254 nm). The column was packed with a strong cation-exchange resin consisting of sulfonated crosslinked styrene/divinylbenzene copolymer in lead (Pb) form. Deionised water was used as a mobile phase at a flowrate of 0.6 mL min-1 . Pyrolysis oils (watersoluble fraction) were diluted in deionised water in a mass ratio of 1:20 and were filtered with 0.2 µm Whatman RC Agilent filter. During analysis temperature of the column was maintained at 70 °C and 10 µL of sample volume was injected into the system at room temperature. The limit of detection for a given species is 0.002 wt. %, while 0.008 wt. % is the limit of quantification. Quantification was performed based on the RID. Analysis of model compound pyrolysis oils was performed on LC-UV (UHPLC 3000 Thermo Fisher Scientific) system, equipped with Kinetex F5 core-shell LC column by phenomenex operated at 35 °C. A mixture of acetonitrile and water was used as eluent (0.2 mL min-1 ), where the concentration of acetonitrile was gradually increased from 10 vol. % to 90 vol. % over 60 min. During the analysis, VWD was set at 254 nm and it was used for the quantification.. 2.4.5. Direct infusion mass spectrometry. The pyrolysis oils (water-soluble fraction) were diluted with deionised water to a concentration of ∼150 mg kg-1 . Ammonium acetate was added to the diluted oil34.

(59) water mixture as an ionisation agent[129,130] . A 1 mL Hamilton syringe loaded onto a syringe pump was used to infuse diluted samples directly into the ESI chamber at a rate of 0.01 mL min-1 . The instrument used was an ESI Ion-Trap mass spectrometer (Bruker amaZon SL, Germany). The ESI MS analysis was accomplished in manual mode using drying temperature of 200 °C, N2 flowrate and nebulizer pressure of 6 L min-1 and 10 psi, respectively. Full scan mass spectra were acquired over the m/z range of 50 - 2000. For MS2 experiments, helium gas was used as a collision gas with a fragmentation amplitude voltage of 1 V and a mass window was 1.5 Da.. 2.4.6. Gel permeation chromatography (GPC). The molecular weight distribution (MW D) of lignin and pyrolysis oils was determined by using GPC (Agilent Technologies 1200) equipped with three columns (7.5 mm x 300 mm, particle size 3 µm) placed in series packed with highly cross-linked polystyrene-divinylbenzene copolymer gel (Varian PLgelMIXED E). The column temperature was maintained at 40 °C during analysis. Samples were dissolved in tetrahydrofuran (THF) in a 1:100 mass ratio and were filtered through a 0.45 µm Whatman RC Agilent filter. THF was used as a mobile phase at a flowrate of 1 mL min-1 . A RID and VWD (λ = 254 nm) were installed. The calibration line was made with 10 polystyrene standards with molecular weights ranging from 162 Da to 29510 Da, which was used for the conversion of elution volume to molecular weight.. 2.4.7. Nuclear magnetic resonance (NMR). 1D NMR: 1 H 1H. NMR spectra were recorded on a Bruker Avance II 600 MHz NMR spectrometer (14.1 T), using a 5 mm triple nucleus TXI 1H- 13C/ 15N/ 2H probe and Avance III 400 MHz NMR spectrometer (9.4 T) provided with a broadband BBFO probe. Both spectrometers are equipped a with z gradient coil producing a maximum gradient strength of 50 gauss cm-1 . 2D NMR: Q-CAHSQC 2D Q-CAHSQC NMR spectra were acquired on a Bruker Avance II 600 MHz spectrometer (14.1 T), using a 5 mm triple nucleus TXI 1H- 13C/ 15N/ 2H probe, equipped at 25 °C without sample spinning using the Q-CAHSQC pulse program[131] . Matrices of 2048 data points for the 1H-dimension and 256 data points for the 13C-dimension were collected with a relaxation delay of 6 s and spectral widths from 14 to -1 ppm and from 35. C2.

No results found